The high level of genetic variability of HIV-1 has presented a major hurdle for AIDS vaccine development. Genetic differences among HIV-1 groups M, N, and O are extensive, ranging from 30% to 50% in gag and env genes, respectively (Gurtler et al, J. Viral. 68:1581-1585 (1994), Vanden Haesevelde et al, J. Viral. 68:1586-1596 (1994), Simon et al, Nat. Med. 4:1032-1037 (1998), Kuiken et al, Human retroviruses and AIDS 2000: a compilation and analysis of nucleic acid and amino acid sequences (Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, N. Mex.)). Viruses within group M are further classified into nine genetically distinct subtypes (A-D, F-H, J and K) (Kuiken et al, Human retroviruses and AIDS 2000: a compilation and analysis of nucleic acid and amino acid sequences (Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, N. Mex., Robertson et al, Science 288:55-56 (2000), Robertson et al, Human retroviruses and AIDS 1999: a compilation and analysis of nucleic acid and amino acid sequences, eds. Kuiken et al (Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, N. Mex.), pp. 492-505 (2000)). With the genetic variation as high as 30% in env genes among HIV-1 subtypes, it has been difficult to consistently elicit cross-subtype T and B cell immune responses against all HIV-1 subtypes. HIV-1 also frequently recombines among different subtypes to create circulating recombinant forms (CRFs) (Robertson et al, Science 288:55-56 (2000), Robertson et al, Human retroviruses and AIDS 1999: a compilation and analysis of nucleic acid and amino acid sequences, eds. Kuiken et al (Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, N. Mex.), pp. 492-505 (2000), Carr et al, Human retroviruses and AIDS 1998: a compilation and analysis of nucleic acid and amino acid sequences, eds. Korber et al (Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, N. Mex.), pp. III-10-III-19 (1998)). Over 20% of HIV-1 isolates are recombinant in geographic areas where multiple subtypes are common (Robertson et al, Nature 374:124-126 (1995), Cornelissen et al, J. virol. 70:8209-8212 (1996), Dowling et al, AIDS 16:1809-1820 (2002)), and high prevalence rates of recombinant viruses may further complicate the design of experimental HIV-1 immunogens.
To overcome these challenges in AIDS vaccine development, three computer models (consensus, ancestor and center of the tree) have been used to generate centralized HIV-1 genes to (Gaschen et al, Science 296:2354-2360 (2002), Gap et al, Science 299:1517-1518 (2003), Nickle et al, Science 299:1515-1517 (2003), Novitsky et al, J. Virol. 76:5435-5451 (2002), Ellenberger et al, Virology 302:155-163 (2002), Korber et al, Science 288:1789-1796 (2000)). The biology of HIV gives rise to star-like phylogenies, and as a consequence of this, the three kinds of sequences differ from each other by 2-5% (Gao et al, Science 299:1517-1518 (2003)). Any of the three centralized gene strategies will reduce the protein distances between immunogens and field virus strains. Consensus sequences minimize the degree of sequence dissimilarity between a vaccine strain and contemporary circulating viruses by creating artificial sequences based on the most common amino acid in each position in an alignment (Gaschen et al, Science 296:2354-2360 (2002)). Ancestral sequences are similar to consensus sequences but are generated using maximum-likelihood phylogenetic analysis methods (Gaschen et al, Science 296:2354-2360 (2002), Nickle et al, Science 299:1515-1517 (2003)). In doing so, this method recreates the hypothetical ancestral genes of the analyzed current wild-type sequences (FIG. 26). Nickle et al proposed another method to generate centralized HIV-1 sequences, center of the tree (COT), that is similar to ancestral sequences but less influenced by outliers (Science 299:1515-1517 (2003)).
The present invention results, at least in part, from the results of studies designed to determine if centralized immunogens can induce both T and B cell immune responses in animals. These studies involved the generation of an artificial group M consensus env gene (CON6), and construction of DNA plasmids and recombinant vaccinia viruses to express CON6 envelopes as soluble gp120 and gp140CF proteins. The results demonstrate that CON6 Env proteins are biologically functional, possess linear, conformational and glycan-dependent epitopes of wild-type HIV-1, and induce cytokine-producing T cells that recognize T cell epitopes of both HIV subtypes B and C. Importantly, CON6 gp120 and gp140CF proteins induce antibodies that neutralize subsets of subtype B and C HIV-1 primary isolates.
The iterative nature of study of the centralized HIV-1 gene approach is derived from the rapidly expanding evolution of HIV-1 sequences, and the fact that sequences collected in the HIV sequence database (that is, the Los Alamos National Database) are continually being updated with new sequences each year. The CON6 gp120 envelope gene derives from Year 1999 Los Alamos National Database sequences, and Con-S derives from Year 2000 Los Alamos National Database sequences. In addition, CON6 has Chinese subtype C V1, V2, V4, and V5 Env sequences, while Con-S has all group M consensus Env constant and variable regions, that have been shortened to minimal-length variable loops. Codon-optimized genes for a series of Year 2003 group M and subtype consensus sequences have been designed, as have a corresponding series of wild-type HIV-1 Env genes for comparison, for use in inducing broadly reactive T and B cell responses to HIV-1 primary isolates.